EVALUATION OF PREFERENCE FOR REINFORCEMENT OR RESPONSE COST CONDITIONS By CRISTINA MARIA WHITEHOUSE A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2010 1
53
Embed
To my Dad, Henry Remsen Whitehouse IIufdcimages.uflib.ufl.edu/UF/E0/04/19/91/00001/whitehouse... · 2013. 5. 31. · & Kahan, 1992; Frentz & Kelly, 1986, Miltenbeger, Parish, Rickert,
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
EVALUATION OF PREFERENCE FOR REINFORCEMENT OR RESPONSE COST CONDITIONS
By
CRISTINA MARIA WHITEHOUSE
A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE
Results and Discussion ........................................................................................... 23
EXPERIMENT II: RESPONSE COST VS. REINFORCEMENT: DIFFERENT SYMBOL SETS ...................................................................................................... 27
Results and Discussion ........................................................................................... 28
EXPERIMENT III: RESPONSE COST VS. REINFORCEMEN: COLOR AND SYMBOLS CONTROLLED ..................................................................................... 33
Table page 3-1. Summary of latencies, point comparisons, and side effects. ................................. 26
4-1. Average latencies for Elena and Allen. .................................................................. 31
4-2. Points at first choice point for Elena and Allen. ...................................................... 32
5-1. Latency summary for all participants. ..................................................................... 40
5-2. Point totals at the first choice point in each phase. ................................................ 41
7
LIST OF FIGURES
Figure page 3-1. Proportion of choice selections per session for Nicole and Maribel. ...................... 25
4-1. Proportion of choice selections per session for Elena and Allen. ........................... 30
5-1. Results for participants showing a preference for reinforcement. .......................... 37
5-2. Results for participants showing indifference. ........................................................ 38
5-3. Results for participant showing a preference for response cost. ............................ 39
6-1. Results for Cole, Kylie, Monica, and Amelia. ......................................................... 45
8
Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science
EVALUATION OF PREFERENCE FOR REINFORCEMENT OR RESPONSE COST
CONDITIONS
By
Cristina Maria Whitehouse
August 2010
Chair: Timothy R. Vollmer Major: Psychology
The use of response cost and reinforcement-based interventions (e.g., token
economies, group level systems, grading with points) is common in academic settings.
Despite the ubiquity of these interventions, few studies have evaluated child preference
for response cost versus reinforcement. Furthermore, the studies have yielded mixed
results. The present study involved assessments of child preference for reinforcement
or response cost in a series of 4 experiments. In each experiment, typically developing
children were repeatedly presented with a computerized matching to sample task under
both reinforcement and response cost conditions. Following exposure to each condition,
children were asked to select their subsequent working conditions. Child selections
were the primary dependent measure of choice. This preparation was repeated using
different stimuli, to assess if preference results could be reproduced. Additionally, this
preparation was repeated using math problems appropriate for the child’s grade level.
Results of Experiments I and II showed that preference was influenced by variables
other than the contingencies presented. In Experiment III, 8 of the 12 participants
showed a preference for reinforcement. However, preliminary results of Experiment IV
indicated that numbering the trials during the session may attenuate any overall
9
10
preference for reinforcement. These data have implications for assessing client
treatment preference and the acceptability of response cost procedures.
CHAPTER 1 INTRODUCTION
When given options, practitioners tend to choose interventions that are effective,
least restrictive, and likely to be implemented by the client or caregivers. Variables that
influence treatment selection decisions include treatment practicality, efficacy, social
acceptability, and client preference (Hanley, Piazza, Fisher, Contrucci, & Maglieri,
1997). Although providing clients with opportunities to choose is an important part of a
client’s right to effective treatment (Bannerman, Sheldon, Sherman, & Harchik, 1990),
research in the area of client treatment preference is sparse.
Recently, however, there have been some investigations that have primarily
focused on client treatment preference. Hanley et al., (1997), for example, examined
preference for two commonly recommended and equally effective interventions,
functional communication training (FCT) and noncontingent reinforcement (NCR), in the
reduction of aggression with two participants. The authors utilized a concurrent-chain
procedure, commonly used in choice experiments, to identify preference. The primary
dependent variable was the number of switch button presses among concurrently
available buttons. Each button was correlated with access to different treatments.
Results for both participants indicated a preference for the FCT training. The authors
suggested that response-dependent schedules may be preferred to response-
independent schedules because participants’ behavior controls the rate of
reinforcement. Importantly, the study established a method for assessing client
treatment preference for children with limited verbal skills.
Hanley, Piazza, Fisher, and Maglieri (2005) applied a similar concurrent-chain
procedure to investigate treatment preference for FCT with and without punishment.
11
Results from the two participants indicated preferences for FCT with punishment. The
punishment component was a 30-s hands-down procedure, and in addition, for one
participant, a 30-s visual block. The authors concluded that if the interventions were
selected based solely on structure or name, FCT with punishment may not have been
chosen because of the punishment component. Furthermore, the authors suggested
that recent trends toward eschewing punishment procedures based on name or
structure would have eliminated a treatment that was both effective and preferred by
these two individuals. The study raises important ethical implications of examining
treatment preference when considering interventions that involve punishment.
The decision to use or not use punishment procedures is a topic of long standing
controversy in behavior analysis. This issue is often debated in the context of
punishment procedures that are highly restrictive or aversive, such as shock. However,
punishment is defined technically as any stimulus change occurring after a response
that decreases the future probability of that response (Michael, 1993). Thus,
interventions such as response cost, which involves contingent loss of reinforcers, are
punishment procedures if they result in reductions in behavior.
Response cost has been shown to be an effective behavior reduction treatment on
a variety of topographies from skin picking (Allen & Harris, 1966) to sleep problems in
children (Piazza & Fisher, 1991). Response cost also is commonly used in academic
environments, such as when points are lost contingent on incorrect answers, and has
been shown to be effective (Phillips, Phillips, Fixsen & Wolf, 1971). Additionally, several
investigations have used academic tasks to compare reinforcement procedures to
Indifferent 6.19 3.79 2.91 7.23 8.15 Mary 6.48 3.84 3.28 8.87 11.45 Anne 6.38 3.54 3.65 9.20 9.35 Preference for Elena 4.35 5.04 3.66 Response Cost 3.37 3.34 3.41
40
Table 5-2. Point totals at the first choice point in each phase. Shaded region = RC Symbols #1 Symbols #2 Math #1 Math #2 Dan 6 2 10 10 9 8 10 8 Callie 6 6 10 10 8 6 9 10
Indifferent 6 10 10 7 Mary 6 9 10 10 Anne 9 9 10 10
Preference for Elena 4 7 Response Cost 6 7
41
JackieCorrespondence
2/8
2/42/4 2/4
2/6
5/8
3/65/12
0.00
0.25
0.50
0.75
1.00
R v RC R v RC #2 R v RC Math R v RC Math #2
Prob
abili
ty
.Selecting R when R was 'fav.'conditionSelecting RC when RC was'f'av.' condition
M aggieCorre s ponde nce
12/12
16/24
6/6 10/1015/1613/14
12/12
5/6
0.00
0.25
0.50
0.75
1.00
R v RC R v RC #2 R v RC M ath R v RC M ath #2
Prob
abili
ty
.
Figure 5-4. Correspondence data for Jackie and Maggie.
42
CHAPTER 6 EXPERIMENT IV: RESPONSE COST VS REINFORCEMENT: NUMBERING TRIALS
Methods
Subjects
Four typically developing children enrolled in afterschool care at a local elementary
school participated in this study. Ages of the children ranged from 7 to 10 years-old.
None of the children had participated in any of the previous experiments.
Procedure
In this experiment, the 10 trials in each component were numbered. Trial numbers
appeared below the number line and to the left of the symbols and math problems.
Additionally, for Amelia, a fifth phase was implemented. This phase was a reversal to
the math problems used in phase 3. All other procedures were identical to Experiment
III.
Design
The design was identical to Experiment III.
Results and Discussion
Figure 6-1 shows the results for the four participants. One participant (Cole,
upper left panel) showed a preference for reinforcement, and one participant (Monica,
lower left panel) initially showed a preference towards reinforcement but that preference
switched to indifference in the last two phases of the experiment. The third participant
(Amelia, lower right panel) showed a preference towards reinforcement but switched to
indifference in the fourth phase. The final participant (Kylie, upper right panel) showed a
pattern of responding not observed in any of the previous experiments. Her preference
43
44
was categorized as inconclusive based on the changes in her preference across the
phases of the experiment.
To address Amelia’s change in preference at during the last 4 sessions of phase
4 of the experiment, a fifth phase was added. The purpose of the phase was to reverse
the math problems used in phase 3 to examine whether the set of math problems was
influencing preference. Amelia continued to show indifference in the reversal to phase 3
math problems.
The preference variability among 3 of the 4 participants suggests that numbering
the trials may temper the apparent preference for reinforcement obtained in Experiment
III. Recall that numbering the trials was added to the current experiment to eliminate
another procedural nuance that appeared to influence preference for either contingency.
However, these results should be viewed cautiously, as only 4 participants have
completed this experiment to date. More participants are necessary in order to make
further comparisons to Experiment III.
Similar to the previous experiments, side effects and differences in latencies
were not obtained. Additionally, point totals at the first choice point did not appear to
influence preference. :
Figure 6-1. Results for Cole, Kylie, Monica, and Amelia
Monica
0.00
0.33
0.67
1.00
0 5 10 15 20 25 30 35
SessionsR einfo rcement
R es po ns e C o s t
Prop
ortio
n of
Cho
ice
Con
ditio
ns
Symbols #1Numbered trials
Symbols #2Numbered Trials
Math #1Numbered Trials
Math #2Numbered T rials
Cole
0.00
0.33
0.67
1.00
0 5 10 15 20 25
Sessions
Reinfo rcement
Res po ns e Co s t
Prop
ortio
n of
Cho
ice
Con
ditio
ns
Symbols #2Numbered Trials
Symbols #1 Numbered trials
Math #1Numbered Trials
Math #2Numbered Trials
Amelia
0.00
0.33
0.67
1.00
0 5 10 15 20 25
Sessions
Symbols #1
Reinfo rcement
Res po ns e Co s tPr
opor
tion
of C
hoic
e C
ondi
tions
Symbols #1Numbered T rials
Symbols #2Numbered Trials
Math #1Numbered T rials
Math #2Numbered T rials
Reversal toMath #1 problems
Kylie
0.00
0.33
0.67
1.00
0 5 10 15 20 25
Sessions
R einfo rcement
R es po ns e Co s t
Prop
ortio
n of
Cho
ice
Con
ditio
ns
umbered trialssymbols #2 Math #1 Math #2
N Numbered Trials Numbered trials Numbered trials
45
CHAPTER 7 GENERAL DISCUSSION
The present experiments evaluated preference for reinforcement or response cost
using simulated academic task and actual academic tasks. Results of Experiments I
and II suggested that any preference for reinforcement or response cost must have
been so slight (if present at all) that it appeared to be outweighed by idiosyncratic
features of the experiment. Thus, all four participants in the first two experiments can be
viewed as relatively indifferent to reinforcement versus response cost. In Experiment III,
there appeared to be a general preference for reinforcement over response cost. In
Experiment IV, when the trials were numbered, the participants showed mixed,
indifferent, and inconclusive preferences. Despite the results from Experiment III, the
overall results collectively suggest that a preference for reinforcement seems marginal
and situational at best. Preferences apparently were controlled by very minor procedural
nuances that were incidental to the experimental arrangement. In Experiments I and II
the participants stated that they preferred certain colors or symbols, respectively, and,
the data supported those statements. In Experiment III, children reported they preferred
reinforcement because it was easier to “keep track” or see how many trials were left to
complete in each component. The data from Experiment IV, though best viewed as
preliminary, suggest that strong overall preference for reinforcement is mitigated when
the participants can easily keep track of the trials in either condition. Altogether, the
findings of these experiments also suggest that the bell and buzzer sounds used in the
Sattler et al. (1976) study, although innocuous at first glance, may have exerted control
over participants’ preferences
46
Side effects were not observed differentially in response cost conditions but were
most often seen in the first component of the first session, regardless of condition.
Furthermore, longer latencies to respond were not associated with response cost as
found in Sattler et al (1976). In all experiments, the mean latencies were minimally
shorter in the child’s preferred condition, which may be due to the child completing more
trials in that condition. For example, if a child’s overall preference was for reinforcement
in a session, 30 trial latencies would be averaged for reinforcement, and 10 trial
latencies from the one forced exposure to response cost condition would be averaged.
Given that the latencies decreased as sessions progressed, it makes sense that the
mean latencies would be lower in the condition with more exposure. In fact, a
retrospective review of data in the forced choice components (thus, controlling for
exposure) showed no discernable difference in mean latency.
Some potential limitations of the current experiments should be noted. First, this
investigation only examined preference for reinforcement or response cost in the
context of an acquisition task. Similar experiments could be arranged to evaluate
problem behavior targeted for reduction. Second, because the same symbols sets were
used in Experiments III and IV, no comparison can be made of accuracy during
reinforcement versus response cost. Additionally, the participants were only given the
option to select either reinforcement or response cost during choice conditions. The
participants did not have an opportunity to select other options such as “I don’t care.” An
evaluation of adding such an option to the experiment is currently ongoing.
There are a number of other potentially interesting future manipulations. Sattler et
al., (1976) suggested that removing points the child had earned may influence
47
preference and increase the probability of observing side effects. Previous response
cost evaluations have utilized removing points earned non-contingently (e.g., Kazdin,
1973; Pace & Foreman, 1982) and points earned contingently (e.g., Phillips, 1968;
Phillips et al., 1971). Though both have been shown to be effective, none of these
studies evaluated child preferences. It is also unclear whether the investigators
observed any side effects, as no discussion of side effects was presented. Thus, an
extension of the current study would be to investigate whether removing earned points
versus noncontingent points influences preference or side effects. Another future
direction would be to manipulate different parameters, such as how many points are
earned or lost (e.g., Kazdin, 1971), and assessing preference. Additionally,
reinforcement and response cost are rarely implemented in isolation. Thus, future
studies should include a combined reinforcement and response cost condition.
Response cost procedures are ubiquitous (e.g., McSweeny, 1978); speeding, if
caught, leads to a fine; failing to pay an electric bill may result in the electricity being
discontinued; and paying a credit card bill late may result in an increase in the annual
percentage rate. Response cost procedures are commonly found in typical classrooms
too. Thus, it is important for researchers to continue to examine such punishment
procedures and to evaluate any preferences and side effects in the context of naturally
occurring instances of reinforcement or response cost.
The issue of client treatment preference becomes especially important if behavior
analysts are to consider such information when making treatment decisions. Outside of
behavior analysis, most studies in the area of client treatment preference assess it
through the use of questionnaires or verbal reports (Dwight-Johnson, Sherbourne, Liao,
48
49
& Wells, 2000). Although Hanley et al. (1997) demonstrated a method to identify
treatment preference with consumers with limited vocal verbal behavior, the use of
client treatment preference assessments remains largely unexplored. Thus, it is
important for behavior analysts to continue to examine the best way to accurately and
efficiently measure preference for consumers with both limited and extensive verbal
behavior. For example, can verbal reports of preference be verified by setting up a
simple concurrent operant treatment selection? Subsequently, how many sessions or
choice points must be completed to identify an overall preference? Presumably,
preference is not static, and a number of known factors (e.g., delay to reinforcement,
amount of reinforcement, effort, severity of the behavior problem) influence choice. For
now, at least in the context of the types of tasks used in the current experiments, it can
be said that any preference for reinforcement over response cost is at best marginal.
LIST OF REFERENCES
Allen, K., & Harris, F. (1966). Elimination of a child's excessive scratching by training the mother in reinforcement procedures. Behaviour Research and Therapy, 4(2), 79-84. doi:10.1016/0005-7967(66)90046-5.
Azrin, N. H. & Holz, W. C. (1966). Punishment. In W.K. Honig (Ed.), Operant behavior:
areas of research and application. East Norwalk, CT US: Appleton-Century-Crofts. Retrieved from PsycINFO database.
Bannerman, D., Sheldon, J., Sherman, J., & Harchik, A. (1990). Balancing the right to
habilitation with the right to personal liberties: The rights of people with developmental disabilities to eat too many doughnuts and take a nap. Journal of Applied Behavior Analysis, 23(1), 79-89. doi:10.1901/jaba.1990.23-79.
Blampied, N., & Kahan, E. (1992). Acceptability of alternative punishments: A
community survey. Behavior Modification, 16(3), 400-413. doi:10.1177/01454455920163006.
Boren, J., & Colman, A. (1970). Some experiments on reinforcement principles within a
psychiatric ward for delinquent soldiers. Journal of Applied Behavior Analysis, 3(1), 29-37. doi:10.1901/jaba.1970.3-29.
Brent, D., & Routh, D. (1978). Response cost and impulsive word recognition errors in
reading-disabled children. Journal of Abnormal Child Psychology: An official publication of the International Society for Research in Child and Adolescent Psychopathology, 6(2), 211-219. doi:10.1007/BF00919126.
Broughton, S., & Lahey, B. (1978). Direct and collateral effects of positive
reinforcement, response cost, and mixed contingencies for academic performance. Journal of School Psychology, 16(2), 126-136. doi:10.1016/0022-4405(78)90051-1.
Carr, J., & Sidener, T. (2002). On the relation between applied behavior analysis and
positive behavioral support. The Behavior Analyst, 25(2), 245-253. Retrieved from PsycINFO database.
Dwight-Johnson, M., Sherbourne, C., Liao, D., & Wells, K. (2000). Treatment
Preferences Among Depressed Primary Care Patients. JGIM: Journal of General Internal Medicine, 15(8), 527-534. doi:10.1046/j.1525-1497.2000.08035.x.
Errickson, E. A., Wyne, M. D., & Routh, D. K. (1973). A response-cost procedure for
reduction of impulsive behavior of academically handicapped children. Journal of Abnormal Child Psychology, 1, 350-357.
50
Fisher, W., Thompson, R., Piazza, C., Crosland, K., & Gotjen, D. (1997). On the relative reinforcing effects of choice and differential consequences. Journal of Applied Behavior Analysis, 30(3), 423-438. doi:10.1901/jaba.1997.30-423.
Florida Administrative Code, Title XXIX, Chapter 393 (2009). Frentz, C., & Kelley, M. (1986). Parents' acceptance of reductive treatment methods:
The influence of problem severity and perception of child behavior. Behavior Therapy, 17(1), 75-81. doi:10.1016/S0005-7894(86)80116-2.
Hanley, G., Piazza, C., Fisher, W., Contrucci, S., & Maglieri, K. (1997). Evaluation of
client preference for function-based treatment packages. Journal of Applied Behavior Analysis, 30(3), 459-473. doi:10.1901/jaba.1997.30-459.
Hanley, G., Piazza, C., Fisher, W., & Maglieri, K. (2005). On The Effectiveness Of And
Preference For Punishment And Extinction Components Of Function-Based Interventions. Journal of Applied Behavior Analysis, 38(1), 51-65. doi:10.1901/jaba.2005.6-04.
Holt, M., & Hobbs, T. (1979). The effects of token reinforcement, feedback and
response cost on standardized test performance. Behaviour Research and Therapy, 17(1), 81-83. doi:10.1016/0005-7967(79)90054-8.
Iwata, B., & Bailey, J. (1974). Reward versus cost token systems: An analysis of the
effects on students and teacher. Journal of Applied Behavior Analysis, 7(4), 567-576. doi:10.1901/jaba.1974.7-567.
Kazdin, A. (1973). The effect of response cost and aversive stimulation in suppressing
punished and nonpunished speech disfluencies. Behavior Therapy, 4(1), 73-82. doi:10.1016/S0005-7894(73)80075-9.
Kazdin, A. (1971). The effect of response cost in suppressing behavior in a pre-
psychotic retardate. Journal of Behavior Therapy and Experimental Psychiatry, 2(2), 137-140. doi:10.1016/0005-7916(71)90029-2.
Latham, G. I. (1995). The Power of Positive Parenting: A Wonderful Way to Raise
Children. Logan, UT: P&T ink. McSweeny, A. (1978). Effects of response cost on the behavior of a million persons:
Charging for directory assistance in Cincinnati. Journal of Applied Behavior Analysis, 11(1), 47-51. doi:10.1901/jaba.1978.11-47.
Michael, J. (1993). Concepts and principles of behavior analysis. Kalamazoo: Society
for the advancement of behavior analysis.
51
52
Miltenberger, R. G., Parrish, J. M., Rickert, V., & Kohr, M. (1989). Assessing treatment acceptability with consumers of outpatient child behavior management services. Child and Family Behavior Therapy, 11, 35-44. doi:10.1300/J019v11n01_03.
Pace, D. M., & Forman, S. G. (1982). Variables related to the effectiveness of response
cost. Psychology in the Schools, 19, 365-370. doi:10.1002/1520-6807(198207)19:3<365::AID-PITS2310190317>3.0.CO;2-T.
Panek, D. M. (1970). Word association learning by chronic schizophrenics on a token
economy ward under conditions of reward and punishment. Journal of Clinical Psychology, 26, 163-167. doi:10.1002/1097-4679(197004)26:2<163::AID-JCLP2270260208>3.0.CO;2-5.
Phillips, E. L. (1968). Achievement place: Token reinforcement procedures in a home-
style rehabilitation setting for "pre-delinquent" boys. Journal of Applied Behavior Analysis, 1, 213-223. doi:10.1901/jaba.1968.1-213.
Phillips, E. L., Phillips, E. A., Fixsen, D. L., & Wolf, M. M. (1971). Achievement place:
Modification of the behaviors of pre-delinquent boys within a token economy. Journal of Applied Behavior Analysis, 4, 45-59. doi:10.1901/jaba.1971.4-45.
Piazza, C. C., & Fisher, W. (1991). A faded bedtime with response cost protocol for
treatment of multiple sleep problems in children. Journal of Applied Behavior Analysis, 24, 129-140. doi:10.1901/jaba.1991.24-129.
Piazza, C. C., Fisher, W. W., Roane, H. S., & Hilker, K. (1999). Predicting and
enhancing the effectiveness of reinforcers and punishers. In A. C. Repp & R. H. Horner (Eds.), Functional analysis of problem behavior (pp. 55-77). Belmont, CA: Wadsworth.
Sattler, H. E., Betz, M. A., & Zellner, R. D. (1978). Children’s preference for response
cost or positive reinforcement as working conditions. The Journal of Psychology, 100, 71-75. Retrieved from PsycINFO database.
Sidman, M. (1989). Coercion and its fallout. Boston, MA: Authors Cooperative, Inc.
Retrieved from PsycINFO database. Sprute, K. A., Williams, R. L., & McLaughlin, T. F. (1990). Effects of a group response
cost contingency procedure on the rate of classroom interruptions with emotionally disturbed secondary students. Child and Family Behavior Therapy, 12, 1-12. doi:10.1300/J019v12n02_01.
BIOGRAPHICAL SKETCH
Cristina Whitehouse graduated from Rollins College in 1995 with a Bachelor of
.Arts. in Psychology. Cristina became interested in Behavior Analysis while at Rollins
and completed her senior thesis on the Personalized System of Instruction with her
undergraduate advisor, Dr. Maria Ruiz. Upon graduation, Cristina began working at
Threshold Inc. providing early intervention behavioral services for children with autism
and developmental delays. Cristina then spent two years working as a Primary
Therapist providing behavioral services at Quest Kids. These opportunities lead to work
at Community Services for Autistic Adults and Children in Rockville, MD where she was
responsible for in-home behavioral programming and staff training.
After moving back to Florida, Cristina spent 3 years working as a Behavior
Analyst providing behavioral services to foster children and teaching foster parents
behavioral parenting skills under the supervision of Dr. Timothy Vollmer. This
opportunity motivated Cristina’s to pursue a graduate degree in applied behavior
analysis, and she began her graduate studies at the University of Florida in 2004. Since
beginning graduate school, Cristina has had the opportunity to conduct research in the
areas of assessment of preference and large scale program evaluation. Cristina will
continue her graduate studies at the University of Florida to obtain her doctoral degree.